Lucene search
+L

5 matches found

osv_ghsa
osv_ghsa
•added 2026/10/28 9:15 p.m.•18 views

GHSA-H3RC-J82P-2C3H vLLM: Untrusted EPS inputs can trigger denial of service or Ghostscript RCE

Summary In vLLM 0.30.0, ImageMediaIO passes attacker-controlled image bytes to Pillow's format autodetection without restricting the accepted formats; the request's MIME label does not constrain decoder selection. An EPS payload labeled image/png reached the host's Ghostscript executable. On a ho...

8.1CVSS7.1AI score0.27974EPSS
SaveExploits7References6
Cvelist
Cvelist
•added 2026/10/26 5:16 p.m.•41 views

CVE-2026-115915 vLLM: Untrusted EPS inputs can trigger denial of service or Ghostscript RCE

vLLM is an inference and serving engine for large language models. Prior to 0.32.0, ImageMediaIO passes attacker-controlled image bytes to Pillow through Image.openBytesIOdata without restricting decoder selection, and the MIME label does not constrain the selected format. Image-capable request...

8.1CVSS0.0066EPSS
SaveExploits0References4
osv_cve
osv_cve
•added 2026/10/26 5:16 p.m.•30 views

CVE-2026-115915 vLLM: Untrusted EPS inputs can trigger denial of service or Ghostscript RCE

vLLM is an inference and serving engine for large language models. Prior to 0.32.0, ImageMediaIO passes attacker-controlled image bytes to Pillow through Image.openBytesIOdata without restricting decoder selection, and the MIME label does not constrain the selected format. Image-capable request...

8.1CVSS6.2AI score
SaveExploits0References6
Vulnrichment
Vulnrichment
•added 2026/10/26 5:16 p.m.•4 views

CVE-2026-115915 vLLM: Untrusted EPS inputs can trigger denial of service or Ghostscript RCE

vLLM is an inference and serving engine for large language models. Prior to 0.32.0, ImageMediaIO passes attacker-controlled image bytes to Pillow through Image.openBytesIOdata without restricting decoder selection, and the MIME label does not constrain the selected format. Image-capable request...

8.1CVSS7.6AI score0.0066EPSS
SaveExploits0References4
CVE
CVE
•added 2026/10/26 5:16 p.m.•77 views

CVE-2026-115915

The vLLM inference engine (prior to version 0.32.0 ) is vulnerable because ImageMediaIO passes attacker-controlled image bytes to Pillow via Image.open(BytesIO(data)) without restricting decoder selection or constraining formats via MIME labels. Attackers can submit specially crafted images, such...

9.8CVSS7.6AI score0.0066EPSS
SaveExploits0References4Affected Software1
Rows per page
Query Builder